Articles

The sex hormone system in carriers of BRCA1/2 mutations:

a case-control study
Martin Widschwendter*, Adam N Rosenthal*, Sue Philpott, Ivana Rizzuto, Lindsay Fraser, Jane Hayward, Maria P Intermaggio,
Christopher K Edlund, Susan J Ramus, Simon A Gayther, Louis Dubeau, Evangelia Ourania Fourkala, Alexey Zaikin, Usha Menon, Ian J Jacobs

Summary
Lancet Oncol 2013; 14: 112632
Published Online
October 17, 2013
http://dx.doi.org/10.1016/
S1470-2045(13)70448-0
See Comment page 1151
Copyright Widschwendter et al.
Open Access article distributed
under the terms of CC BY
*Contributed equally
Department of Womens
Cancer, UCL EGA Institute for
Womens Health
(Prof M Widschwendter MD,
A N Rosenthal PhD,
S Philpott MSc, I Rizzuto MD,
L Fraser BSc, J Hayward MSc,
E O Fourkala PhD,
Prof A Zaikin PhD,
Prof U Menon MD,
Prof I J Jacobs MD), and
Department of Mathematics
(Prof A Zaikin PhD), University
College London, London, UK;
Barts Cancer Institute, Barts and
the London School of Medicine
and Dentistry, London, UK
(A N Rosenthal); Department of
Preventive Medicine
(M P Intermaggio MSc,
C K Edlund MSc, S J Ramus PhD,
Prof S A Gayther PhD), and
Department of Pathology
(Prof L Dubeau PhD), Keck School
of Medicine, USC/Norris
Comprehensive Cancer Center,
University of Southern
California, Los Angeles, CA, USA;
and Faculty of Medical and
Human Sciences, University of
Manchester and Manchester
Academic Health Science Centre,
Manchester, UK (Prof I J Jacobs)
Correspondence to:
Prof Martin Widschwendter,
Department of Womens Cancer,
UCL EGA Institute for Womens
Health, University College
London, London WC1E 6AU, UK
M.Widschwendter@ucl.ac.uk

For more on the UK FOCSS

see http://www.
instituteforwomenshealth.ucl.
ac.uk/academic_research/
gynaecologicalcancer

1226

Background Penetrance for breast cancer, ovarian cancer, or both in carriers of BRCA1/BRCA2 mutations is
disproportionately high. Sex hormone dysregulation and altered end-organ hormone sensitivity might explain this
organ-specic penetrance. We sought to identify dierences in hormone regulation between carriers of BRCA1/2 and
women who are negative for BRCA1/2 mutations.
Methods We assessed endometrial thickness for each menstrual cycle day (as an index of hormone regulation) in
393 scans from 228 women in the UK Familial Ovarian Cancer Screening Study (UK FOCSS) known to carry either
mutation and 1573 scans from 754 women known to be negative for the mutations. To quantify dierences in
endometrial thickness we focused on days 1014 and days 2126, and calculated the area under the curve. We then
compared serum oestradiol and progesterone titres during these days of the menstrual cycle in the same groups.
Follicular and luteal oestradiol and progesterone serum titres were grouped into quartiles and odds ratios were
calculated with logistic regression.
Findings Follicular phase endometrial thickness of carriers of the mutations adjusted for age and day of the menstrual
cycle was higher (odds ratio [OR] 111, 95% CI 103120; p=00063) and luteal phase endometrial thickness lower
(090, 083098; p=0027) than for women negative for the mutations. Median luteal phase titres of progesterone
were 121% higher (p=000037) in carriers than in women negative for the mutations, and for oestradiol were 33%
higher (p=0007)ie, 59% of carriers had concentrations of serum progesterone that would have been in the top
quartile of concentrations in the control group (OR 80, 95% CI 215257; p=0008).
Interpretation Carriers of BRCA1/BRCA2 mutations are exposed to higher titres of oestradiol and progesterone
known risk-factors for breast cancer. Higher titres of oestradiol in carriers are compatible with this hormone having a
role in ovarian carcinogenesis in such women. Our ndings could not be explained by dierential contraceptive pill use.
Funding Eve Appeal, European Union, Cancer Research UK, and US National Institutes of Health.

Introduction
In all inherited cancer syndromes the germline mutation
is thought to have a so-called local eect in an organ that
is predisposed to the development of cancer, because
these mutations do not cause cancers in all organs. For
example, the increased cancer risk in carriers of the
BRCA1 and BRCA2 mutations is predominantly that of
breast cancer, ovarian cancer, or both.1 These mutations
are thought to cause cancer via a defect in DNA damage
response or in the DNA repair pathway.2 However, this
defect does not explain the organ-specic cancer
penetrance. That removal of both ovaries and Fallopian
tubes reduces not only the risk of ovarian but also breast
cancer3 implicates a systemic dysregulation of hormone
production in carriers of the mutation, which aects both
the Mllerian epithelium, as the cell of origin for ovarian
cancer,4 and breast epithelium. Evidence from preclinical
models suggests that both hormone production and
hormone sensitivity of end organs is altered in carriers of
the BRCA1 mutation. Studies in animals57 showed that
mice carrying a Brca1 mutation in the steroid-hormoneproducing granulosa cells had a longer pro-oestrus phase,
corresponding with the oestrogen-dominant follicular

phase of the human menstrual cycle. Furthermore,

serum oestradiol titres in mutant mice were higher than
those of wild-type mice when both groups were stimulated
with exogenous gonadotropins. Also, the insulin-like
growth factor system has a fundamental role in
endometrial biology, acting via autocrine and paracrine
mechanisms.8 There are strong interactions between the
insulin-like growth factor and BRCA1 signalling
pathways, which also involve oestrogen signalling;9
hence, it is possible that the endometrium of carriers of
the BRCA1 mutation has altered sensitivity to hormones.
Cyclical change in oestradiol and progesterone titres
alters endometrial thickness and menstrual bleeding in
premenopausal women. We postulate that both endometrial thickness as a functional index of hormonal
activity in a target organ that undergoes cyclic changes,
and serum ovarian steroid hormone titres at particular
stages of the menstrual cycle, would dier between
carriers of the BRCA1/BRCA2 mutations and women
who are known to have no BRCA1/BRCA2 mutations.
The UK Familial Ovarian Cancer Screening Study
(UK FOCSS; registered with Current Controlled Trials,
number ISRCTN32794457)10 has accrued sucient
www.thelancet.com/oncology Vol 14 November 2013

Articles

prospective data and samples to allow assessment of our

hypotheses. Because many women in the study had
undergone clinical genetic testing for BRCA1/2
mutations, we had a cohort of women known to carry the
mutation and a cohort known to be negative for the
mutation to compare. However, most of the women in
the study had not undergone clinical genetics testing.
To establish the mutation status of these women, highthroughput next generation sequencing provided the
fastest and most cost-eective method of rapidly
detecting carriers of the mutations, albeit with potentially
lower sensitivity than clinical testing, which in the UK
uses a combination of Sanger sequencing (including
limited Ashkenazi mutation screening where
appropriate) and multiplex ligation probe amplication.

Methods
Participants
After ethical approval (Eastern MREC 97/5/007),
UKFOCSS recruited from 44 UK regional centres.
Between June, 2002, and September, 2010, women older
than 35 years, at an estimated minimum 10% lifetime
risk of ovarian cancer based on family history or
predisposing germline gene mutations (appendix) were
recruited and screening data and outcomes were
collected prospectively. So far, about 25% of the study
population have undergone clinically initiated testing for
BRCA1 and BRCA2 mutations (either before or after
recruitment).
After ethical approval (Joint University College London/
University College London Hospital Ethics Committee,
ref 06/Q0505/102), we selected all premenopausal
UKFOCSS participants with no previous or subsequent
history of cancer (to avoid including women on hormonal
therapy or women with a subclinical cancer, which could
have triggered an altered hormonal environment and
which could have confounded our analyses) and no
intrauterine device.

Procedures
Ovarian cancer screening was done with transvaginal
sonography to assess ovarian morphology and serum
CA125 tumour marker measurement. All centres
scanned study participants and all scans were done by
sonographers, radiologists, or gynaecologists approved
by the UKs National Health Service (NHS), employed by
the local hospital for gynaecological scanning, and
subject to local NHS quality control. Initially, CA125 tests
were done annually, but after 2007 these tests were done
every 4 months. Samples were taken in EDTA-containing
tubes at womens primary care practices and posted to
our laboratory for aliquoting, CA125 testing, and storage
at 80C. Samples were discarded if they reached the
laboratory more than 56 h after venepuncture or were
haemolysed. Endometrial thickness and date of last
menstrual period were routinely recorded during
sonography. Information on use of oral contraceptive
www.thelancet.com/oncology Vol 14 November 2013

pills was collected in a general health questionnaire sent

to all UK FOCSS participants in 2011, asking whether
they had used the oral contraceptive pill in each of their
third, fourth, and fth decades.
We assessed UK FOCSS volunteers for germline
mutations in the coding sequence of BRCA1 and BRCA2
with the Illumina HiSeq2000 sequencer (San Diego, CA
USA; appendix). Participants were included if they
matched our inclusion criteria and had provided a DNA
sample.
Oestradiol and progesterone were measured with
automated immunoassays on the Elecsys 2010 analyser
(Roche Diagnostics GmbH, Mannheim, Germany). The
measuring range for oestradiol was 1815 800 pmol/L
and for progesterone was 0095191 nmol/L. Standard
Westgard rules were applied. Mutation carriers and
controls were randomly mixed in batches and analysis
was masked. Single lot numbers of reagent and calibrator
were used. The oestradiol intra-assay coecient of
variability (CV) is 1657% and interassay CV is
2362%. The progesterone intra-assay CV is 1527%
and interassay CV is 3754%.

Statistical analysis

See Online for appendix

Endometrial thickness for each group was averaged for

each day of menstrual cycle and smoothed using the
averaging running window of 5 days. For example, if ve
women had samples from cycle day 10, we rst took the
mean result for these ve women. We then took the
mean of the means for days 812 and plotted this as the
value for day 10. On start and end days of the menstrual
cycle the dependencies were prolongedeg, for day 1,
endometrial thickness was averaged over days 30, 31, 1, 2,
and 3. To quantify dierences we focused on days 1014
and days 2126 (because these were the times in the
cycle when dierences were most pronounced), and
calculated the area under the curve (AUC). Dierences
in AUC and in distributions represented in boxplots
were estimated with the Mann-Whitney U test. To
construct the distributions for AUCs, the endometrial
thickness values were bootstrapped: for each cycle day
the thickness value was sampled with replacement, then
the obtained dependence was averaged using a window
of 5 days and the AUC was calculated for the obtained
dependence between days 1014 and 2126. The odds
ratios (ORs) for BRCA status were obtained for
endometrial thickness, treated as a continuous predictor
variable, and adjusted in the logistic regression for age
and the specic menstrual cycle day as continuous
variables.
Follicular and luteal oestradiol and progesterone serum
titres were grouped into quartiles and ORs were
calculated with logistic regression. Because endometrial
thickness, oestradiol, and progesterone data did not
dier signicantly between carriers of BRCA1 and
BRCA2 mutations (appendix), we combined the data of
the mutation carriers to increase statistical power with
1227

Articles

Results
2808 scans from 1460 UK FOCSS volunteers for whom
ET available and who do not have cancer

BRCA1 and BRCA2 testing done

1966 scans in 982 volunteers

1573 scans in 754 negative for both mutations
203 scans in 116 carrying the BRCA1 mutation
190 scans in 112 carrying the BRCA2 mutation

Mutation status known during trial (A)

429 scans in 254 volunteers
86 scans in 51 negative for both mutations
177 scans in 104 carrying the BRCA1 mutation
166 scans in 99 carrying the BRCA2 mutation

Mutation status assessed after trial (B)

1537 scans in 728 volunteers
1487 scans in 703 negative for both mutations
26 scans in 12 carrying the BRCA1 mutation
24 scans in 13 carrying the BRCA2 mutation

Figure 1: Availability of transvaginal ultrasound scans

Scans from women within the UK Familial Ovarian Cancer Screening Study (UK FOCSS) who had clinically initiated
BRCA testing before or during the trial (A) or who had BRCA testing by next generation sequencing after the trial
nished (B). In the UK, clinically initiated testing is done primarily on women with stronger family histories and
aected living relatives. Therefore the only dierence between B and A is that the former would have had a lower
a-priori risk of being mutation carriersthis explains the lower prevalence of mutations recorded in B versus A.
However, there was no dierence in ultrasound methods, data collection, sample collection, or storage for these
women, so this is not a source of bias. Consequently no statistical adjustment was deemed necessary.
ET=endometrial thickness.

BRCA1 and
BRCA2
negative

BRCA1 or
BRCA2
mutation

p value*

1099 (77%)

219 (78%)

061

332 (23%)

61 (22%)

All scans
No contraceptive pill use
Contraceptive pill use

Mutation status known during the trial

No contraceptive pill use

46 (81%)

191 (79%)

Contraceptive pill use

11 (19%)

52 (21%)

073

Mutation status assessed after the trial

No contraceptive pill use
Contraceptive pill use

1053 (77%)

28 (76%)

321 (23%)

9 (24%)

089

Use of oral contraceptive pills in the same decade as endometrial thickness scan was
available for 87% of the 1966 endometrial thickness scans done. *From test.
Women whose mutation status was established with clinical genetics testing.
Women whose mutation status was established with next generation sequencing.

Table 1: Proportions of endometrial thickness scans with known

contraceptive pill use during decade of ultrasound scan

the limited number of samples available. ORs for BRCA

status were obtained for endometrial thickness, treated
as a continuous predictor variable, and adjusted for age
and the specic menstrual cycle day. Software R version
2.11.1 was used for statistics.

Role of the funding source

The sponsor of the study had no role in study design,
data collection, data analysis, data interpretation, or
writing of the report. The corresponding author had full
access to all the data in the study and had nal
responsibility for the decision to submit for publication.
1228

2808 endometrial-thickness measurements and associated

last menstrual period dates were available from 1460
eligible women, including 1573 endometrial thickness
measurements in women negative for both mutations,
203 in carriers of BCRA1, and 190 in carriers of BRCA2
(baseline data shown in the appendix). 254 women had
BRCA testing done during the course of the UK FOCSS
and 728 women (who had provided DNA) had their BRCA
status assessed after the trial by next generation sequencing
(gure 1). Information on use of oral contraceptive pills
during the decade when the transvaginal scan was done
was available for 1711 (87%) of 1966 endometrial thickness
measurements and was not signicantly dierent in any
of the groups assessed (table 1).
409 stored serum samples from cycle days 1014 and
days 2126 were available from known carriers of BRCA1
(n=38) and BRCA2 (n=32) mutations and negative
controls (n=339) for oestradiol and progesterone testing.
Figure 2 and table 2 show endometrial thickness in the
study groups by menstrual cycle day. Endometrial
thickness was higher in the follicular phase in carriers
than in controls negative for the mutations, but lower
than that of controls in the luteal phase. The AUC of
endometrial thickness in both follicular and luteal phase
diered signicantly when comparing carriers of either
mutation with those who were carriers of neither
mutation (p<00001 for both comparisons) and
endometrial thickness results did not dier signicantly
between carriers of the mutations (appendix). In view of
this statistically similar endometrial thickness pattern in
carriers of the mutations, we combined these groups,
and noted clear dierences between the combined carrier
group and the group negative for mutations (gure 2 and
table 2). In the follicular phase, the carrier groups
endometrial thickness was signicantly higher, whereas
in the luteal phase endometrial thickness was
signicantly lower than that of non-carriers, even after
using logistic regression analysis and adjusting for age
and menstrual cycle day. To assess whether a womans
knowledge of her mutation status might aect the results
(eg, by changing her lifestyle in an attempt to minimise
her cancer risk), we separately analysed the 728 women
in the next generation sequencing group who did not
know their mutation status during the trial. The same
endometrial thickness pattern between mutation carriers
and non-mutation carriers was noted (appendix). To be
certain that the endometrial thickness dierences were
not due to dierential oral contraceptive pill use, we
analysed the 1318 endometrial thickness scans for which
the women had reported no oral contraceptive pill use in
the decade the scan was done (table 1). Again, we noted
the same endometrial thickness patterns as before
(appendix), with higher follicular phase endometrial
thickness and lower luteal phase endometrial thickness
in carriers of the mutations than for women negative for
the mutations.
www.thelancet.com/oncology Vol 14 November 2013

Articles

Discussion
Our ndings show clear dierences in cyclical
endometrial thickness in carriers of BRCA1/2 mutations
compared with wild-type controls. We have also shown
congruent changes in sex-hormone titres, which are one
of the probable explanations for this dierence. Our
cohort of well matched premenopausal carriers and
controls with combined endometrial thickness information, and serum samples with known menstrual cycle
data, was ideal for testing the hypothesis that the organspecic cancer penetrance in carriers is due to hormonal
dysregulation. To our knowledge, the UK FOCSS
population is the largest cohort for whom this
information is available (panel).
Our ndings of high luteal phase progesterone titres in
carriers of BRCA mutations, associated with decreased
endometrial thickness, are in complete concordance with
each other, and might relate to a defect in steroid-hormone
regulation in carriers that results in an end-organ eect.
We speculate that the high luteal phase oestradiol we
recorded in carriers triggers increased expression of
progesterone receptors,11 thus potentiating any possible
mutagenic eect of the higher luteal progesterone. Our
ndings support those from studies in mice carrying a
Brca1 mutation in ovarian granulosa cells.6,7 Although data
on endogenous premenopausal progesterone exposure
www.thelancet.com/oncology Vol 14 November 2013

BRCA1 and 2 negative

BRCA1 mutation
BRCA2 mutation

10

Endometrial thickness (mm)

B
BRCA1 and 2 negative
BRCA1 and 2 mutation

10

Endometrial thickness (mm)

The dierences in endometrial thickness between

carriers and those negative for the mutations could be a
consequence of dierent hormonal sensitivity of the
target tissue (ie, the endometrium) in carriers of the
mutations, triggered by dierent titres of the steroid
hormones known to regulate endometrial biology, or a
combination of the two. Although it is dicult to assess
hormonal sensitivity directly, to assess the triggering
threshold we analysed oestradiol and progesterone in
stored serum samples from all premenopausal carriers
of the mutations who provided samples between days
1014 (follicular phase) and 2126 (luteal phase; n=59,
mean age 406 years, 70 samples), and all women
negative for the mutations (n=283, mean age 435 years,
339 samples). Again, there were no dierences between
carriers of either mutation (appendix). Oestradiol titres
did not dier signicantly between the carrier and
mutation-negative groups during days 1014 (appendix).
Progesterone titres during day 1014 were not measured
in all women because pilot data in 38 carriers and
44 controls showed such low titres (median 127 nmol/L
in carriers and controls) that no signicant dierences
would become apparent on testing all available samples.
Median luteal phase titres of progesterone were 121%
higher (p=000034) in carriers than in women negative
for the mutations, and oestradiol titres were 33% higher
(p=0007)ie, 59% of carriers had concentrations of
serum progesterone that would have been in the top
quartile of concentrations in the control group (OR 80,
95% CI 215257; p=0008; appendix, gure 3).

5
OR=111* (95% CI 103120)
p=00063

OR=090* (95% CI 083098)

p=0027

4
0

10

15
Menstrual cycle (days)

20

25

30

Figure 2: Endometrial thickness as a function of the menstrual cycle

(A) Endometrial thickness calculated from 1573 transvaginal ultrasound scans from 754 women negative for both
mutations, 203 scans from 116 carriers of the BRCA1 mutation, and 190 scans from 112 carriers of the BRCA2
mutation. (B) Endometrial thickness in 754 women negative for both mutations (1573 scans) and the combined
228 women who were carriers of either mutation (393 scans). OR=odds ratio. *Adjusted for menstrual cycle day
and age.

and cancer risk is less extensive and less conclusive,12

postmenopausal exogenous progesterone exposure is a
well established risk factor for breast cancer.11,1315 We were
recently part of a collaborative report suggesting that
progestogens cause breast cancer by inducing expression
of the receptor activator of NF-B ligand (RANKL), and
that deleting the receptor for this ligand delays the onset
of progestogen-driven breast cancers.16 There is already an
1229

Articles

Area under curve

p value

Days 1014
of cycle

Days 2126
of cycle

AUC

AUC

40

32

31

43

BRCA2 mutation

42

32

32

41

BRCA1/2 negative

308

28

274

45

BRCA1/2 mutated

82

32

63

41

BRCA1/2 negative

308

28

274

45

<00001

Separate BRCA analysis

BRCA1 mutation

<00001

Combined BRCA analysis

AUC=area under curve. n=number of volunteers.

Table 2: Endometrial thickness as a function of menstrual cycle in

women by BRCA1 and BRCA2 mutation status

p=00074

80

1000

Progesterone (nmol/L)

Oestradiol (pmol/L)

800

600

400

p=000034

60

40

20

0
Negative

BRCA1/2
mutation

Negative

BRCA1/2
mutation

Figure 3: Serum progesterone and oestradiol analysis during luteal phase of

the menstrual cycle
Boxplots with horizontal line show 25%, 50%, and 75% quartiles. Error bars
show 95% CIs.

antibody against RANKL (denosumab), which is used at

present for the treatment of osteoporosis. Our ndings
might therefore provide the basis for trials of this antibody
as a chemopreventive agent for BRCA1/2-triggered breast
cancer. In conjunction with data that show the potential of
a progesterone antagonist to prevent BRCA1-mediated
mammary tumorigenesis in mice,17 our ndings provide
an additional rationale for treatment that interferes with
progesterone signalling to prevent breast cancer in
carriers of BRCA1/BRCA2 mutations. One such
compound (the progesterone receptor modulator,
ulipristal acetate) has been used successfully for treatment
of uterine broids, with fewer side-eects than agonists of
luteinising-hormone-releasing hormone.18,19 We speculate
that such drugs could be used as chemoprophylaxis for
women at high risk of breast cancer.
The recorded higher follicular phase endometrial
thickness in carriers of the BRCA1/2 mutations was not
associated with any dierences in follicular phase titres
of oestradiol and progesterone between carriers and
1230

Systematic review
We searched PubMed for studies on BRCA mutation and
alterations in the female premenopausal reproductive
hormone system published in English between Sept 1, 1993,
and Feb 28, 2013. We used the search terms (BRCA1 or
BRCA2) and (endocrine or hormones or endometrium
or menstrual). We established that studies in animals
showed aberrant hormonal regulation upon loss of BRCA1 in
granulosa cells and that premenopausal surgical resection of
the ovaries in women carrying the BRCA1 or BRCA2 mutation
led to a substantial reduction in the risk of breast cancer.3
Systemic endocrine eects of BRCA1 or BRCA2 mutations
have not previously been assessed in humans.
Interpretation
Our ndings suggest that BRCA1/2 germline mutations are
driving carcinogenesis only in part via altered molecular
pathways (eg, those involved in DNA repair) in the organ at
risk, and that BRCA1/2-associated changes in the endocrine
system are additional factors. These insights could act as a
major impetus for novel chemoprevention trials using
strategies that can exploit the hormonal dysregulation in
carriers of BRCA1/2 mutations. Potential agents for these
trials include selective oestrogen or progesterone receptor
modulators, and the anti-RANKL (receptor activator of NF-B
ligand) antibody denosumab (currently used for osteoporosis
treatment, but known to block the downstream carcinogenic
eects of progestogens).

200

Panel: Research in context

controls. We speculate that the higher follicular phase

endometrial thickness in carriers might be a consequence
of altered endometrial sensitivity to steroid hormones.
Evidence on endometrial cancer penetrance in carriers20
suggested that excess endometrial cancer risk was
explained by use of tamoxifen. This nding suggests that
the endometrium of carriers might be more sensitive to
oestrogen-receptor agonists.
Oestrogen replacement usage21 and obesity (a hyperoestrogenic state)22 are established epidemiological risk
factors for ovarian cancer, implicating hormonal
dysregulation in its pathogenesis. Our data, suggesting
higher oestradiol titres in the luteal phase of the
menstrual cycle in women who are carriers of the
BRCA1/2 mutations compared with women negative for
the mutations supports this hypothesis. We speculate
that the higher titres of progesterone with concordant
reduced endometrial thickness recorded in the luteal
phase in the carrier group compared with the control
group might explain why the penetrance for the third
hormonally triggered cancernamely endometrial
cancerin carriers is much lower than that for ovarian
and breast cancer; it is well recognised that progestogens
suppress endometrial proliferation,23 resulting in a
thinner endometrial thickness and lower lifetime risk of
endometrial cancer.24
www.thelancet.com/oncology Vol 14 November 2013

Articles

Our study has certain limitations. Although we

excluded women with a levonorgestrel-releasing
intrauterine device, we had limited information on use of
oral contraceptive pills. We only recorded use of the pill
in the same decade as the endometrial scan. However,
these data were available on 87% of the women for whom
we had endometrial thickness data. Based on these data,
oral contraceptive use during the same decade as the
ultrasound scan did not dier between carriers and
controls. Furthermore, in view of the age of the cohort
(all were >35 years) and their known increased risk of
breast cancer, it is probable the proportion using the
contraceptive pill at the time of sample donation would
have been even lower than the data above. Most
importantly, excluding women with unreported pill use
and those known to have taken the pill in the same
decade as the scan did not reduce the statistical
signicance of the endometrial thickness dierences
between carriers and controls.
The control group could have included participants
that carry BRCA1/2 mutations who were missed using
our next generation sequencing mutation-detection
methods (eg, mutations in regions of low sequence
coverage or large genomic rearrangements). Although it
is not possible to estimate the frequency of missed
mutations, the individuals screened were unaected and
therefore had at most a 50% chance of inheriting a
germline mutation even if it were present in their family.
Therefore, the frequency of BRCA1/BRCA2-positive
individuals in the screen-negative group is probably very
low. Any mutation carriers in the screen-negative group
would bias our nding towards the null, suggesting that
our study has underestimated rather than overestimated
the strength of the recorded eects.
We had insucient numbers of scans with last
menstrual period dates of a cycle length greater than
28 days to draw any conclusions about whether carriers
of BRCA1/BRCA2 mutations might have a longer
menstrual cycle than women negative for the mutations,
as was evident in the previously described mouse
model.57 If this were the case, then we speculate that
longer cumulative exposure, and the observed higher
absolute titres of progesterone and to a lesser extent
oestradiol, might contribute to the excess breast cancer
risk of carriers of the mutations.
Our study cannot address whether endometrial
thickness and hormone titres are respectively a marker
and eector of breast cancer risk within a BRCA1/2mutant population. Women with such a mutation have a
very high (up to 85%) lifetime risk of breast cancer.
Therefore many women undergo risk-reducing salpingooophorectomy, mastectomy, or both. As a result
addressing the question as to whether altered endometrial
thickness and steroid hormone titres in premenopausal
women are markers for subsequent breast cancer risk or
not would be dicult to do. It would need a substantial
prospective long-term study of premenopausal BRCAwww.thelancet.com/oncology Vol 14 November 2013

carriers who were unwilling to undergo risk-reducing

surgery but willing to be followed up for decades. Our
ndings suggest that sex steroids are one of the major
drivers for development of breast cancer in this
population.
We deliberately excluded patients who had a previous
diagnosis of breast or ovarian cancer because they could
have been on antiendocrine therapy or chemotherapy,
which would have altered their ovarian function and
biased our ndings. We also excluded the few women
who developed breast or ovarian cancer subsequent to
their scan or serum sample donation. It is well known
that sex steroids are locally produced within invasive but
also within non-invasive breast carcinoma,25 which might
not have been clinically apparent at the time of sample
donation. Furthermore, small subclinical ovarian
tumours might potentially interfere with normal
hormone production. Hence we decided a priori not to
include women with a diagnosis of breast or ovarian
cancer subsequent to sample donation to avoid any bias
that would favour the hypothesis of an aberrant endocrine
system being associated with cancer development in
high-risk women. It could be argued that by excluding
cases of subsequent breast cancer, we have diluted the
breast cancer risk prole of our study group. However,
despite this, we have noted hormonal changes that would
be expected to increase breast cancer risk (ie, raised
oestradiol and progesterone in carriers vs controls).
Furthermore, because of the young age of the study
group (all were premenopausal), most have not yet
reached the age at which their breast cancer would be
likely to occur, so most of the excess breast cancer risk of
the study population has yet to accrue. Excluding the
small number of women who we know developed breast
cancer subsequent to sample donation would be likely to
minimise dierences between carriers and controls,
making our ndings more compelling.
To improve statistical power, we combined carriers of
BRCA1 and BRCA2 mutations into a single study group.
Although we acknowledge that it is possible that there
are biological dierences between these groups relevant
to our investigation, we did not identify statistical
dierences between them with respect to endometrial
thickness or hormone titres (appendix). Furthermore,
gure 2 clearly shows that both groups have a similar
pattern of higher follicular phase endometrial thickness
and lower luteal phase endometrial thickness compared
with women negative for the mutations.
Although we have in eect analysed multiple crosssectional data rather than longitudinal data, we have still
identied statistically signicant dierences between
carriers and controls. Furthermore, longitudinal data
would need a prospective study of a similar number of
mutation carriers and controls, requiring daily
venepuncture and transvaginal sonography. We feel such
a study would be logistically challenging and recruitment
for it would be extremely dicult.
1231

Articles

The control group was not a random sample from the

general population, but rather was deliberately selected
for high familial risk of ovarian cancer (to minimise any
dierences between mutation carriers and controls).
The control groups results might not therefore be
representative of the general population. However, their
shared increased ovarian cancer risk makes them the
most appropriate control group to compare with carriers
of BRCA1/BRCA2 mutations.
In conclusion, our ndings provide novel insights into
the high penetrance for breast cancer (via higher
progesterone and oestrogen titres) and also possibly
ovarian cancer (via higher oestrogen titres and potentially
lower titres of anti-Mllerian hormone26) in carriers of
BRCA1/BRCA2 mutations. Our ndings provide a
rationale for novel cancer prevention strategies in these
women. They also provide a possible explanation for the
absence (via higher progesterone titres) of high
penetrance for endometrial cancer in carriers.
Contributors
MW had the idea and together with ANR did the literature search, data
collection, data analysis, and data interpretation; designed the study;
prepared the gures; and wrote the report. SP, IR, LF, EOF, UM, and
IJJ collected, analysed, and interpreted the data. LD interpreted the data.
JH, MPI, CKE, SJR, and SAG did the next generation sequencing.
AZ did the statistics and generated gures and table.
Conicts of interest
We declare that we have no conicts of interest.
For more on the Eve Appeal see
http://www.eveappeal.org.uk/

Acknowledgments
This work was funded by the Eve Appeal and the European Unions
Seventh Framework Programme (FP7/2007-2013) under grant
agreement number 305428 (Project EpiFemCare) and done at UCLH/
UCL, which received a proportion of its funding from the Department of
Health NIHR Biomedical Research Centres (BRC) funding scheme. The
UK FOCSS was core funded by Cancer Research UK (grants C315/A2621
and C1005/A6383), the UKs Department of Health, and the Eve Appeal
with additional support from grants MRC8804/A/A7058, CA152990, and
CA086381 from the NCI Early Detection Research Network. Next
generation sequencing was done by the USC Molecular Genomic Core
and was supported in part by award number P30CA014089 from the
National Cancer Institute. The content is solely the responsibility of the
authors and does not necessarily represent the ocial views of the
National Cancer Institute or the National Institutes of Health.
LD received grant support from US National Institutes (grant numbers
R01 CA133117 and RO1 CA119078). IJJ is an NIHR Senior Investigator.
We thank David Conti for his help in the bioinformatic analysis of next
generation sequencing data.
References
1
King MC, Marks JH, Mandell JB. Breast and ovarian cancer risks
due to inherited mutations in BRCA1 and BRCA2. Science 2003;
302: 64346.
2
Venkitaraman AR. Cancer susceptibility and the functions of
BRCA1 and BRCA2. Cell 2002; 108: 17182.
3
Domchek SM, Friebel TM, Singer CF, et al. Association of
risk-reducing surgery in BRCA1 or BRCA2 mutation carriers with
cancer risk and mortality. JAMA 2010; 304: 96775.
4
Dubeau L. The cell of origin of ovarian epithelial tumours.
Lancet Oncol 2008; 9: 119197.
5
Chodankar R, Kwang S, Sangiorgi F, et al. Cell-nonautonomous
induction of ovarian and uterine serous cystadenomas in mice
lacking a functional Brca1 in ovarian granulosa cells. Curr Biol 2005;
15: 56165.

1232

9
10

11

12

13
14

15

16

17

18

19

20

21

22

23

24

25

26

Hong H, Yen HY, Brockmeyer A, et al. Changes in the mouse

estrus cycle in response to Brca1 inactivation suggest a potential
link between risk factors for familial and sporadic ovarian cancer.
Cancer Res 2010; 70: 22128.
Yen HY, Gabet Y, Liu Y, et al. Alterations in Brca1 expression in
mouse ovarian granulosa cells have short-term and long-term
consequences on estrogen-responsive organs. Lab Invest 2012;
92: 80211.
Zhou J, Dsupin BA, Giudice LC, Bondy CA. Insulin-like growth
factor system gene expression in human endometrium during the
menstrual cycle. J Clin Endocrinol Metab 1994; 79: 172334.
Werner H, Bruchim I. IGF-1 and BRCA1 signalling pathways in
familial cancer. Lancet Oncol 2012; 13: e53744.
Rosenthal AN, Fraser L, Manchanda R, et al. Results of annual
screening in phase I of the United Kingdom familial ovarian cancer
screening study highlight the need for strict adherence to screening
schedule. J Clin Oncol 2013; 31: 4957.
Brisken C. Progesterone signalling in breast cancer: a neglected
hormone coming into the limelight. Nat Rev Cancer 2013;
13: 38596.
Endogenous Hormones and Breast Cancer Collaborative Group.
Sex hormones and risk of breast cancer in premenopausal women:
a collaborative reanalysis of individual participant data from seven
prospective studies. Lancet Oncol 2013; 14: 100919.
Beral V. Breast cancer and hormone-replacement therapy in the
Million Women Study. Lancet 2003; 362: 41927.
Chlebowski RT, Kuller LH, Prentice RL, et al. Breast cancer after
use of estrogen plus progestin in postmenopausal women.
N Engl J Med 2009; 360: 57387.
Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benets of
estrogen plus progestin in healthy postmenopausal women:
principal results From the Womens Health Initiative randomized
controlled trial. JAMA 2002; 288: 32133.
Schramek D, Leibbrandt A, Sigl V, et al. Osteoclast dierentiation
factor RANKL controls development of progestin-driven mammary
cancer. Nature 2010; 468: 98102.
Poole AJ, Li Y, Kim Y, Lin SC, Lee WH, Lee EY. Prevention of
Brca1-mediated mammary tumorigenesis in mice by a progesterone
antagonist. Science 2006; 314: 146770.
Donnez J, Tomaszewski J, Vazquez F, et al. Ulipristal acetate versus
leuprolide acetate for uterine broids. N Engl J Med 2012;
366: 42132.
Donnez J, Tatarchuk TF, Bouchard P, et al. Ulipristal acetate versus
placebo for broid treatment before surgery. N Engl J Med 2012;
366: 40920.
Segev Y, Iqbal J, Lubinski J, et al. The incidence of endometrial
cancer in women with BRCA1 and BRCA2 mutations: an
international prospective cohort study. Gynecol Oncol 2013;
130: 12731.
Beral V, Bull D, Green J, Reeves G. Ovarian cancer and hormone
replacement therapy in the Million Women Study. Lancet 2007;
369: 170310.
Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D. Cancer
incidence and mortality in relation to body mass index in the
Million Women Study: cohort study. BMJ 2007; 335: 1134.
Li Q, Kannan A, DeMayo FJ, et al. The antiproliferative action of
progesterone in uterine epithelium is mediated by Hand2. Science
2011; 331: 91216.
Yang S, Thiel KW, Leslie KK. Progesterone: the ultimate
endometrial tumor suppressor. Trends Endocrinol Metab 2011;
22: 14552.
Suzuki T, Miki Y, Ohuchi N, Sasano H. Intratumoral estrogen
production in breast carcinoma: signicance of aromatase.
Breast Cancer 2008; 15: 27077.
Titus S, Li F, Stobezki R, et al. Impairment of BRCA1-related DNA
double-strand break repair leads to ovarian aging in mice and
humans. Sci Transl Med 2013; 5: 172ra21.

www.thelancet.com/oncology Vol 14 November 2013